[Babel-users] atomic updates and ipv6 cache flushing issues

[Juliusz Chroboczek]

> > 2. When was NLM_F_REPLACE introduced?
> 
> I think we have been using it on Olsr.org for years... so the feature
> was already present in kernel 2.4.

That's excellent news.  Count me suprised.

My git tells me the #define was introduced in 2012, but that's just
moving the code from include/linux to uapi/include/linux.  There are
occurrences in net/ipv4/fib_trie.c from 2006, so we're probably safe.

Thanks for the hint, Henning.  All that remains is for somebody to
explain to me what's the deal with flushing the route cache.

-- Juliusz